mmarco_v2_hi_train / README.md
Sean MacAvaney
commit files to HF hub
a63e5ec
---
pretty_name: '`mmarco/v2/hi/train`'
viewer: false
source_datasets: ['irds/mmarco_v2_hi']
task_categories:
- text-retrieval
---
# Dataset Card for `mmarco/v2/hi/train`
The `mmarco/v2/hi/train` dataset, provided by the [ir-datasets](https://ir-datasets.com/) package.
For more information about the dataset, see the [documentation](https://ir-datasets.com/mmarco#mmarco/v2/hi/train).
# Data
This dataset provides:
- `queries` (i.e., topics); count=808,731
- `qrels`: (relevance assessments); count=532,761
- `docpairs`; count=39,780,811
- For `docs`, use [`irds/mmarco_v2_hi`](https://huggingface.co/datasets/irds/mmarco_v2_hi)
## Usage
```python
from datasets import load_dataset
queries = load_dataset('irds/mmarco_v2_hi_train', 'queries')
for record in queries:
record # {'query_id': ..., 'text': ...}
qrels = load_dataset('irds/mmarco_v2_hi_train', 'qrels')
for record in qrels:
record # {'query_id': ..., 'doc_id': ..., 'relevance': ..., 'iteration': ...}
docpairs = load_dataset('irds/mmarco_v2_hi_train', 'docpairs')
for record in docpairs:
record # {'query_id': ..., 'doc_id_a': ..., 'doc_id_b': ...}
```
Note that calling `load_dataset` will download the dataset (or provide access instructions when it's not public) and make a copy of the
data in 🤗 Dataset format.
## Citation Information
```
@article{Bonifacio2021MMarco,
title={{mMARCO}: A Multilingual Version of {MS MARCO} Passage Ranking Dataset},
author={Luiz Henrique Bonifacio and Israel Campiotti and Roberto Lotufo and Rodrigo Nogueira},
year={2021},
journal={arXiv:2108.13897}
}
```